Post on 14-Apr-2018
7/29/2019 Ump Fulltext01
1/62
ACTA
UNIVERSITATIS
UPSALIENSIS
UPPSALA
2013
Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology1029
Measurement and modellingof unbalanced magnetic pull
MATTIAS WALLIN
ISSN 1651-6214
ISBN 978-91-554-8619-8urn:nbn:se:uu:diva-196490
in hydropower generators
7/29/2019 Ump Fulltext01
2/62
Dissertation presented at Uppsala University to be publicly examined in Polhemssalen,ngstrm Laboratory, Lgerhyddsvgen 1, Uppsala, Friday, April 26, 2013 at 13:15 for thedegree of Doctor of Philosophy. The examination will be conducted in English.
Abstract
Wallin, M. 2013. Measurement and modelling of unbalanced magnetic pull in hydropower
generators. Acta Universitatis Upsaliensis. Digital Comprehensive Summaries ofUppsala Dissertations from the Faculty of Science and Technology1029. 51 pp. Uppsala.ISBN 978-91-554-8619-8.
Hydropower research is often perceived to be an old and exhausted field of study but with ageingequipment and the need for more intermittent operation caused by an increased share of otherrenewable energy sources new challenges lie ahead.
The main focus of this dissertation are the electromagnetic forces resulting from nonuniformair gap flux, whether it be caused by rotor eccentricity or a faulty field winding. Resultsare predominantly obtained from measurements on an experimental generator and numericalsimulations.
With the computational capacity available today it is possible to numerically analyse physicalphenomena that previously could only be studied with analytical tools. Numerical models canalso be expanded to encompass more than one aspect of generator operation in coupled field-circuit models without model complexity surpassing computer capability.
Three studies of unbalanced magnetic pull, UMP, in synchronous salient pole generatorsconstitute the main part of this thesis.
The first is a study of how parallel stator circuits affect the unbalanced magnetic pull caused byrotor eccentricity. Depending on the relationship between the geometry of the separate circuitsand the direction of the eccentricity it was found that parallel circuits could reduce the UMPsubstantially.
Secondly, an investigation of the effect of damper winding configuration on UMP wasperformed. The results showed that damper winding resistivity and the distance between the
damper bars in a pole determine the effectiveness of the damper winding in reducing the UMP.Simulations of a production machine indicate that the reduction can be substantial from damperwindings with low resistivity.
The third study analyses the consequences of field winding interturn short circuits. Apartfrom a resulting rotating unbalanced magnetic pull it is found that the unaffected poles with thesame polarity as the affected pole experience an increase in flux density.
In a fourth article a new stand still frequency response, SSFR, test method includingmeasurements of damper winding voltage and current is presented. It is found that the identifiedmodels are capable of predicting the stator to damper transfer function both with and withoutthe damper winding measurements included.
Keywords: Damper winding, eddy currents, field winding short circuit, finite element
method, hydropower generator, parallel circuits, rotor eccentricity, salient poles, synchronousgenerators, synchronous machines, UMP, unbalanced magnetic pull.
Mattias Wallin, Uppsala University, Department of Engineering Sciences, Electricity,Box 534, SE-751 21 Uppsala, Sweden.
Mattias Wallin 2013
ISSN 1651-6214ISBN 978-91-554-8619-8urn:nbn:se:uu:diva-196490 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-196490)
7/29/2019 Ump Fulltext01
3/62
Till Ann
7/29/2019 Ump Fulltext01
4/62
7/29/2019 Ump Fulltext01
5/62
List of papers
This thesis is based on the following papers, which are referred to in the textby their roman numerals.
I Wallin, M., Ranlf M. and Lundin U., Design and construction of asynchronous generator test setup, XIX International Conference on
Electrical Machines, September 2010.
II Wallin, M., Ranlf, M. and Lundin, U., Reduction of unbalancedmagnetic pull in synchronous machines due to parallel circuits, IEEETransactions on Magnetics, vol. 47, no. 12, pp. 4827-4833, 2011.
III Wallin, M., Bladh, J. and Lundin, U., Damper winding influence onunbalanced magnetic pull in synchronous machines with rotoreccentricity, submitted to IEEE Transactions on Magnetics,November 2012.
IV Wallin, M., and Lundin, U., Dynamic unbalanced pull from fieldwinding turn short circuit, submitted to Electric Power Componentsand Systems, February 2013.
V Bladh, J., Wallin, M., Saarinen, L. and Lundin, U., Stand-stillfrequency response test including damper winding measurements,unpublished manuscript.
Reprints were made with permission from the publishers.
7/29/2019 Ump Fulltext01
6/62
7/29/2019 Ump Fulltext01
7/62
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Hydropower in a new light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Project background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Content of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Synchronous generators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.1 Rotor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.2 Stator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3 Magnetic circuit equivalent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3 Eccentricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73.1 Permeance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83.2 Unbalanced magnetic pull . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
4 Finite element analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114.1 The field-circuit problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114.2 Geometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124.3 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
5 Research areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155.1 Parallel circuits and rotor eccentricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155.2 Damper windings and rotor eccentricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175.3 Field winding interturn short circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195.4 Standstill frequency response test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
6 Experimental generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216.1 Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216.2 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
7 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317.1 Magnetic saturation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317.2 Parallel circuits and rotor eccentricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317.3 Damper windings and rotor eccentricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347.4 Field winding interturn short circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357.5 Results not included in the articles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
8 Summary of articles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
9 Svensk sammanfatting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
10 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
7/29/2019 Ump Fulltext01
8/62
7/29/2019 Ump Fulltext01
9/62
Nomenclature
AcronymsAcronym Description
C Continuous damper windingD13-C Continuous damper winding with bars 1 and 3D13-NC Noncontinuous damper winding with bars 1 and 3D2-C Continuous damper winding with bar 2FE Finite elementFEM Finite element methodLAN Local area networkMMF Magnetomotive forceNC Noncontinuous damper winding
PC Parallel circuitsSSFR Standstill frequency responseTHD Total harmonic distortionUMP Unbalanced magnetic pullWD Without damper windingWLAN Wireless local area network
Greek symbolsSymbol Unit Description
m Air gap length - Relative eccentricity0 Vs/Am Permeability of empty spacer - Relative permeability rad Mechanical frequency
Wb Magnetic flux N/m2 Stresss m Stator slot pitch rad Angular air gap position
7/29/2019 Ump Fulltext01
10/62
Latin symbolsSymbol Unit Description
a, b, c - Electrical phasesA Wb/m Magnetic vector potential
Az Wb/m z-component of the magnetic vector potentialBn T Normal component of the magnetic flux densityBt T Tangential component of the magnetic flux densityd m Rotor displacementFr N Radial forceFt N Tangential force
I A CurrentIf A Field currentMf At Field winding magnetomotive force
Mi At Pole magnetomotive forceN - Neutral point in a three phase Y windingRg A/Wb Air gap reluctanceRr A/Wb Rotor reluctance
Rr m Rotor radiusRs A/Wb Stator reluctanceTm Nm Mechanical torqueSr m Stator radiusV V Voltage
7/29/2019 Ump Fulltext01
11/62
1. Introduction
In 1878 Aristide Bergs called hydroelectric power the white coal, Fr. lahouille blanche, to communicate the abundance and availability of the energystored in the ice and snow of the French Alps.
Hydropower on an industrial scale in Sweden started in 1910 with the com-missioning of the first four generators of the Olidan station in Gta lv in thesouth west of Sweden. The first large hydropower station in northern Sweden,Porjus, started to deliver electricity in 1915.
Until the introduction of nuclear power to the Swedish electricity produc-tion in the seventies and eighties hydropower accounted for the majority ofthe electricity used in Sweden. It was, and still is, vital to the Swedish welfareand industry. Today hydropower accounts for 45 percent, or 65 TWh, of theelectricity used in Sweden.1
1.1 Hydropower in a new light
Even though the technology is mature a greater understanding of hydropowergenerators is needed to avoid failures and undue shortening of the expectedservice life of the machines. Malfunctioning hydropower generators can leadboth to uncertainty in the electricity supply and to increased electricity prices,affecting both households and industries negatively.
Beside the economic impact of nonoperating generators there are environ-mental gains to make from a high degree of utilisation of our existing hydro-power stations.
It is also possible to study electrical machines in new ways due to the com-
putational power available today. Simulations of more complex models wherefinite element magnetic field simulations are integrated with models of thegenerators electric circuits, mechanical properties and thermal characteristicshas resulted in more accurate representations of the generators.
1.2 Project backgroundUnbalanced magnetic pull, UMP, occurs when the rotor is not perfectly centered
in the stator or the magnetic flux density in the air gap of a generator is notrotationally symmetric due to some other cause.
1A third of the energy used in Sweden is in the form of electricity [1].
1
7/29/2019 Ump Fulltext01
12/62
UMP can shorten service life and cause failure through increased wear onbearings and other support structures of the rotor and the stator. Asymmet-ric air gap flux also induces damper winding and intrastator currents whichgive rise to increased resistive losses and a higher rate of thermal ageing. This
thesis investigates three aspects of unbalanced magnetic pull in hydropowergenerators: how parallel stator circuits can reduce UMP caused by rotor ec-centricity, the effects of damper windings under rotor eccentricity and fieldwinding interturn short circuits as a cause of UMP
The work on which the thesis is based has been carried out at the Division ofelectricity at Uppsala University. Results have been obtained from numericalsimulations, analytical models and measurements on the experimental gener-ator built by the author. The generator was named Svante after the donor ofthe motor upon which the experimental generator was built.
The research presented in this thesis is part of an Elektra programme calledDevelopment of electromagnetic load models for simulation of hydropowerrotor-generator systems. Elektras sponsors are ABB, the Swedish TransportAdministration (Trafikverket), Swedish National Grid (Svenska Kraftnt) and,through Swedenergy (Svensk energi), several electricity production and trans-mission companies. Elektras aim is to improve the competitiveness of theSwedish industry through increased research collaboration between universit-ies and the industry.
1.3 Content of the thesisIn the first chapter after the introduction some synchronous generator basicsare presented. Chapter 3 explains the geometry of rotor eccentricity. In chapter4, comes a description of the finite element method, FEM, used in the nu-merical simulations. After that follows a chapter describing the four researchareas: parallel circuits and rotor eccentricity, damper windings and rotor ec-centricity, field winding interturn short circuits and a standstill frequency re-
sponse testing. Chapter 6 describes the mechanical construction considera-tions behind the design of the experimental generator. Information about theinstalled sensors and available measurements signals is also presented. Res-ults are presented in chapter 7. The dissertation ends with a brief summary ofthe included articles.
2
7/29/2019 Ump Fulltext01
13/62
2. Synchronous generators
This chapter is an introduction to synchronous generators. It also containssome terminology used in the remainder of the thesis and the articles.
Electricity can be generated in several ways, for instance chemically in bat-teries and fuel cells, through the photovoltaic effect in solar cells or from elec-tromagnetic induction in generators. The most common type of generators are
rotating synchronous generators. Focus is placed on salient pole synchronousgenerators since these are the most common hydroelectric generators but muchof the material is applicable to round rotor generators as well as synchronousmotors.
At the most basic level hydropower generators convert the potential energystored in a body of water to electrical energy through the intermediate stepof rotating mechanical energy. The torque on the turbine shaft caused by thepressure or kinetic energy of the water drives the rotor shown in figure 2.1. Therotating magnetic field from the field winding on the rotor induces a voltage
in the stator winding which in turn will drive a current through any load con-nected to the generator. Below follows a description of the most importantcomponents of a hydropower generator. The names and positions of somegenerator components are shown in figure 2.1.
2.1 Rotor
A shaft usually connects the turbine runner directly to the rotor and the wholeassembly is supported by one thrust and two or more radial fluid film bearings.The mass of the rotating components, excluding the force from the water, canbe in excess of 1000 tonnes on a large generator. This rotating mass contrib-utes to the inertia and frequency stability of the grid.
Rotor material
To minimise eddy currents from oscillatory transients and stator slot ripple theentire rotor, apart from the shaft, is made of laminated steel. Electrical sheetsteel with a high relative permeability is used to maximise the flux for a givenfield current.
3
7/29/2019 Ump Fulltext01
14/62
Field windingElectromagnetic induction requires a conducting loop and a time magneticfield. On wound field, as opposed to permanent magnet, machines it is the fieldwinding which generates the magnetic flux required to create the generator
voltage. The field windings of the separate poles are often connected in series.At every other pole the field winding changes direction around the pole bodyso the pole fluxes alternate between positive and negative or north and south.
An increase in either the number of field winding turns or the field currentincreases the air gap flux density, B, and hence the total flux that generates thestator voltage. The current density and resistive losses, i.e. heat generation, inthe field winding puts an upper limit on the field current. Magnetic saturationof the rotor and stator steel in combination with the upper limit of the fieldcurrent governs the stator voltage obtainable.
Stator
Stator slot
Stator coil
Stator tooth
Rotor
Pole body
Pole shoe
Damper bar
Air gap
Field winding
Figure 2.1. A schematic drawing of a wound field synchronous generator with six
poles.
Damper windingDamper windings are axially oriented metal bars, usually copper, placed closeto the air gap in the pole shoes of synchronous generators used to dampenoscillatory, electrical or mechanical, transients and to reduce the associated
asynchronous air gap flux. Two forms of damper winding connections arecommon. In the first configurations only the damper bars within a pole areconnected and the only path for any interpole damper currents are through thecontact between the damper bars and the iron components of the rotor. The
4
7/29/2019 Ump Fulltext01
15/62
other alternative is a continuous damper winding where the damper bars ofeach pole are directly connected to the bars of the adjacent poles.
Voltages are induced in the damper bars not only from transients but alsodue to the stator slots during steady state operation. Losses in the damper bars
from stator slot ripple are reduced by keeping the damper slot pitch close tothe stator slot pitch, s [2].To minimise stator voltage harmonics and problems they can cause on the
power grid, the damper bars can be tangentially displaced relative to the centreline of the pole, alternating the directions of displacement [3]. Asymmetric-ally positioned damper bars can be seen in figure 2.1. A shift of the entirepole, instead of just the damper bars, is another common method. Contraryto damper losses stator harmonics can also be reduced by keeping the damperslot pitch close to 0.8s [2].
The effects of damper windings on UMP and the currents induced in thedamper winding from rotor eccentricity are studied in III .
2.2 StatorTogether with the air gap and the rotor the stator forms the magnetic circuit ofthe generator. Electrical sheet steel with a high permeance is used to maximisethe flux in the stator core. The flux in the stator changes direction with theelectrical frequency and the steel is laminated, usually at a thickness of 0.5 or0.35 millimetres, to minimise eddy currents.
The alternating flux and resistive losses in the stator winding generates heatand the stator core is equipped with air ducts for cooling. Water cooled gener-ators also exist.
Stator windingAs the alternating flux from the rotor poles passes the coils of the stator a
voltage is induced in the coils. The winding is normally a Y connected threephase winding. If a load is connected to the generator this voltage will drive acurrent through the stator winding and the load.
2.3 Magnetic circuit equivalentDepending on the aim of the analysis there are several ways of modelling elec-trical machines and one simple representation is through a magnetic circuit. In
article IV this method is used to investigate the eff
ects of an asymmetric fieldwinding.In figure 2.2 Mi is the magnetomotive force, MMF, caused by the field
winding of each pole. Mi it is a function of the number of turns of the field
5
7/29/2019 Ump Fulltext01
16/62
Rr
M6
Rg
B6
Rs
Rs
Rr
M10
Rg
B10
Rs
Rr
M11
Rg
B11
Rs
Rr
M12
Rg
B12
Rs
Rr
M5
Rg
B5
Figure 2.2. Magnetic circuit model of the experiment generator. The MMF for poleeleven,M11, is in article IV given a lower value compared to the other poles.
winding the pole and the field current. Rr, Rg andRs are the reluctances of therotor, air gap and stator respectively. Reluctance is the magnetic resistance, orequivalently, the inverse of the permeance. Bi is the flux density in the air gap.
6
7/29/2019 Ump Fulltext01
17/62
3. Eccentricity
Articles II and III investigate different aspects of rotor eccentricity and gen-erator performance. Rotor eccentricity gives rise to unbalanced magnetic pullwhich puts unwanted stress on rotor and stator supports. It can also causeexcessive wear on the rotor bearings. Below follows, as a background to thediscussion about UMP in articles II and III , a description of the geometry ofrotor eccentricity.
Ideally in a generator the rotor and the stator are perfectly concentric but
wear and manufacturing tolerances will inevitably lead to some degree of rotoreccentricity. Other forms of air gap asymmetries, such as non-parallel statorand rotor axes or a non-circular stator, can also occur but this study is limitedto rotor eccentricity.
An example of rotor eccentricity is shown in figure 3.1, where the variablesused in this chapter also are defined. Rotor eccentricity arises when the statorand rotor centres, Os and Or in figure 3.1, do not coincide.
If the displacement dand the direction of displacement are constant, andnot a function of time, the eccentricity is said to be static and the air gap
length is a function of the angular position round the air gap only, = f(). Ina dynamic eccentricity the centre of the rotor is not fixed and the air gap lengthis a function of both position and time = f(, t). The investigations in articlesII and III and the discussion below are limited to static eccentricity. The airgap length, (), as defined in figure 3.1 is, at any position , for a given fixed
() =Rs (L1(,)+L2(,)) (3.1)
where Rs is the stator radius and the other symbols are defined in figure 3.1.L
1can be written as
L1 = dcos(). (3.2)
and the distance L2 is equal to
L2 =
R2rL
23 =
R2r (dsin())2. (3.3)
where Rr is the rotor radius. Normally d will be small relative to Rr and L2will be close to Rr. IfRr is substituted for L2 the air gap can be written as
() =Rs (L1(,)+Rr) (3.4)
which can be rewritten as
() = 0(1 cos()) (3.5)
7
7/29/2019 Ump Fulltext01
18/62
(,)
0
Rr
x
y
dOs
OrL1
L2
L3
RotorStator
Figure 3.1. An example of a non-centred rotor. Os and Or are the axial centres of thestator and rotor respectively. The dashed circle shows the position of a centered rotor.0 is the nominal air gap. The displacement of the rotor relative to the stator is d andthe direction of the displacement is .
where 0 is the nominal air gap length and is the relative eccentricity definedas the displacement divided by the nominal air gap,
=d
0. (3.6)
3.1 Permeance
The ability of the air gap to conduct magnetic flux, its permeance, is
() =0
(). (3.7)
where 0 is the permeability of empty space and air is assumed to have arelative permeability, r, of one. If() is substituted with equation (3.5) can be written as
() =0
0
1
1 cos(). (3.8)
The reluctance of the rotor and stator will cause the total permeance of thegenerator, , to be lower than . At saturated levels of flux density the rotorand stator components of will also be a function of. The magnitude of the
8
7/29/2019 Ump Fulltext01
19/62
permeance and its variation round the air gap causes the unbalanced magneticpull discussed in articles II and III and it is the rate of change in flux density,which can be modelled by the change in permeance, that drives the damperwinding currents in article III.
3.2 Unbalanced magnetic pullThe air gap flux density, B, is a function of the MMF caused by the fieldwinding, Mf, and the total permeance of the generator . Mf is a function ofrotor position, i.e. time. However, in a multipole machine it is the length of theair gap that to a large extent will determine the size and direction of the UMPcaused by a certain displacement and Mf will hence be treated as a constant
below.In a machine with static eccentricity the direction of displacement, , isconstant and below it has for simplicity been assumed to be zero. The normal,i.e. radial, and tangential magnetic stresses in the air gap are
n =B2n()B
2t()
20and t=
Bn()Bt()
0(3.9)
where n and t indicates normal and tangential components. At no load it isthe normal component which is of interest and with the assumption that Bt is
small relative to Bn equation (3.9a) becomes
n =B2n()
20(3.10)
If it is further assumed that the permeability of the steel is infinite and hencethe permeance of the generator is equal to the air gap permeance the expressionfor Bn in the equation above can be written
Bn() = ()Mf =
0
0
Mf
1 cos() . (3.11)
Insertion of this expression for Bn into equation (3.10) results in
=2M2
f
20=0M
2f
220
1
(1 cos())2(3.12)
where the index n has been omitted.From the stress the force components in the x and y directions can be cal-
culated as
Fx =
2
0
RL cosd (3.13)
9
7/29/2019 Ump Fulltext01
20/62
and
Fy =
2
0
RL sind (3.14)
where R is the average radius of the air gap and L is the length of the generator.Depending on whether the function for Mf and the relative eccentricity is
known or the air gap flux density is known from measurements the total unbal-anced magnetic pull for a generator can be calculated from these equations.
10
7/29/2019 Ump Fulltext01
21/62
4. Finite element analysis
Articles II , III and IV investigated the unbalanced magnetic pull caused by ro-tor eccentricity and field winding short circuits. The effectofdifferent machineconfigurations on the UMP was also studied. Parallel to the measurementsmagnetic field simulations of the experimental generator and generators usedin hydropower production were performed using the finite element method.
Four types of materials are used in the simulations: electrical steel, copper,insulation and air. The steel has a variable relative permeability to allow for
magnetic saturation. It is also possible to assign different steel qualities to forexample the rotor and the stator.
4.1 The field-circuit problemIn the simulations the coupled problems of the magnetic field and the rotor,stator and damper winding electrical circuits are solved simultaneously [4,5].
Finite element problemThe finite element problem for a generator is set up to solve Ampres lawfor the magnetic vector potential, A, in the domain enclosed by the boundaryconditions, given the applied field winding current. Because the problem canbe represented by a 2D problem A only has one component, Az, perpendicularto the xy plane.
Between the nodes of the elements in the mesh the solution is found throughinterpolation of the base functions. These functions can be linear quadratic orof higher order. The work presented here mostly employs linear base functionsbut quadratic functions have been used for comparisons.
Circuit equationsThe magnetic flux from the solution of the finite element problem inducesa voltage in the damper and stator windings of the generator. Through thecircuit equations for the stator and damper windings the resulting currents
in the windings are calculated. These currents aff
ect the flux density and inthe coupled problem the circuit equations are solved together with the finiteelement problem in an iterative process. When a solution is found for a timestep the solver proceeds to the next time step.
11
7/29/2019 Ump Fulltext01
22/62
4.2 Geometry2D finite element analysis was used in the articles mentioned above. Twobenefits of 2D simulations compared to 3D simulations is the reduction incomputational time and the relative ease with which the geometry can be im-
plemented.In simulations of a machine with a symmetric air gap flux density the geo-
metry can often be reduced to one or two poles, equivalent to a half or a fullelectrical cycle. Fractional slot windings might require a larger part of themachine to be modelled. When a non-uniform air gap or a machine with anotherwise asymmetric air gap flux density is studied simulation of the full geo-metry is required.
Mesh and boundary conditionsThe principle of the finite element method is to divide the analysed area into alarge number of small elements and to find a solution for the field equation foreach element. All elements taken together form the mesh. Figure 4.1 containsa section of the generator geometry and the mesh. To avoid remeshing atevery time step during dynamic simulations the rotor and stator are meshedindividually and a sliding mesh interface is used between them. A uniformmesh element size is used in the air gap to allow for the sliding mesh. In figure
4.1 it can also be seen that the mesh element sizes vary and are smaller wherethe geometry contains complex detail.
Figure 4.1. A section of the air gap. One of the poles with its three damper bars canbe seen to the left in the pictures. The stator with its coils make up the right half of thepicture. The elements on each side of the middle of the air gap have equal tangentiallength to allow for the sliding mesh.
12
7/29/2019 Ump Fulltext01
23/62
The boundaries of the finite element problem are shown in figure 4.2. Thearea studied is limited by the inside of the rotor ring as the inner boundaryand the stator back as the outer boundary. Outside this domain the material,air, has a low permeability relative to the steel in the generator and the flux
density is assumed to be zero. The effect of this is that no flux can escape thearea enclosed by the boundary conditions or, equivalently, at the boundary thehomogeneous Dirichlet condition Az = 0 states that the field lines has to beparallel to the boundary.
The interface of the sliding meshes of the rotor and the stator are shown bythe blue line in figure 4.2.
When a section of the generator is simulated it is common to put periodicboundary conditions on the section boundaries that are parallel to a radiustaken from the machine centre. Periodic boundary conditions are not used in
the simulations of a full machine geometry presented here.
Homogenous Dirichletcondition, Az = 0
Sliding mesh
Figure 4.2. The boundaries of the finite element solution. The red circles representhomogeneous Dirichlet conditions, Az = 0, while the blue line indicates the slidingboundary between the rotor and the stator meshes.
Three-dimensional effectsPhenomena which are overlooked in 2D analysis are for example three dimen-sional effects such as eddy currents in the rotor and stator steel, nonlinearities
13
7/29/2019 Ump Fulltext01
24/62
in the magnetic flux and the flux leakage impedance caused by air ducts [6]and end effects, e.g. end flux leakage and coil end inductances. End effectsincrease in importance when the machine has a large diameter relative to itslength. Hydropower generators commonly have a high diameter to length ra-
tio compared to turbo generators and end effects should not be ignored whenstudying the former. Coil end inductances can for example be included in thecoupled circuit described in section 4.1.
EccentricityThe eccentricity used in articles II and III was implemented with the variablepermeance method described in [7]. Instead of changing the air gap length,, in the calculation of the permeance, =0r/, the relative permeability,
r, previously assumed to be one, is changed round the air gap to obtain theequivalent effect on the flux density. With this method it is possible to simulatedynamic eccentricities at other frequencies than the synchronous as well asirregular air gaps without remeshing the geometry at every time step.
4.3 SolutionsAfter the solution for Az has converged, either in a static or a dynamic simu-
lation, post-processing is used to extract more information from the solution.The magnetic pull and the torque are calculated through the Maxwell stresstensor from the normal and tangential flux densities in the air gap. For thetangential flux density an average is taken over part of the air gap in a methodsimilar to the one proposed in [8]. The voltages and currents in the generatorwindings are also calculated.
14
7/29/2019 Ump Fulltext01
25/62
5. Research areas
5.1 Parallel circuits and rotor eccentricityMultipole synchronous generators are often equipped with parallel circuits inthe stator winding to reduce current density and resistive losses. In article IIthe effect of parallel stator circuits on the UMP caused by rotor eccentricity isstudied. The experiment required the stator winding to be divided into two sep-
arate circuits. In this study both FE simulations and an analytical model basedon a varying permeance in the air gap were used together with the measure-ments. A relative eccentricity of 24 percent was used in both measurementsand simulations.
ConfigurationsFigure 5.1 contains an overview of the two stator winding configurations usedin the experimental generator. The division is placed parallel to the x axis.
Hydropower generators will often have more than two circuits and each circuitcan be distributed over more than one section of the stator. Phase a will be usedin the discussion below but phases b and c are treated analogously. When thetwo circuits are connected in series, per phase, the result is a single circuitstator winding as shown in figures 5.1(a) and 5.2(a).
With points a1 and a2 unconnected the generator has two separate circuitsand it is possible to measure the two circuit voltages separately. In the config-urations with two circuits on the experimental generator the terminal voltagewill, for any given field current, be half of that of the single circuit configura-
tion. When the circuits are connected in parallel the total current drawn by aload can be twice as high without an increase of the current density, and res-istive losses, in the stator coils. The measurements and simulations were per-formed with a relative eccentricity of 0.24, a field current of 15A and withoutthe damper winding mounted.
Parallel circuit voltage
As shown in chapter 3 the non-uniform permeance caused by the rotor ec-centricity will result in a varying flux density round the circumference of thegenerator. This variation in flux density can, if the stator winding consists oftwo or more parallel circuits, induce different voltages in the stator circuits.
15
7/29/2019 Ump Fulltext01
26/62
N
u1C a
N1
N2
uPC,1
uPC,2
a1
a2
y
x
y
x
Figure 5.1. With each phase winding divided into two sections it is possible to connectthe sections either in series or in parallel. N stands for neutral. Phases b and c are notdepicted in the graph.
The stator voltage is induced by the varying flux in the air gap and it istherefore, at a given rotational speed, proportional to the flux density shown inequation (3.11). Displacement of the rotor in the direction y, = 90 in figure3.1, will result in the largest flux density difference, and generate the largestvoltage difference, between the two circuits. A first order Taylor approxima-tion and integration of equation (3.11) shows the effect of the eccentricity onthe parallel circuit voltage
u K
ba
1
1 cos()d K
ba
1+ cos() d (5.1)
where K is a proportionality constant. If the integral is taken for the two halvesof figure 5.1(b), i.e. from 0 to 180 degrees, and from 180 to 360 degrees theresults are
uPC,1 K
0
1+ cos(/2) d= K(+2) (5.2)
and
uPC,2 K
2
1+ cos(/2) d= K(2) (5.3)
where the integration intervals are given in radians instead of degrees. Fromequations (5.2) and (5.3) it can be seen that when the rotor is displaced in the ydirection the voltage difference between the two circuits is proportional to thedegree of displacement, .
16
7/29/2019 Ump Fulltext01
27/62
N
ua1 ia
ua2
ua
ub1
ibub2
uc1
icuc2
N
ia2ua2
ua
ia1ua1
ib2
ub ib1 ic2
ic1
uc
Figure 5.2. The two possible stator configurations. In the configuration to the right,ia1 and ia2 must be equal during no-load operation.
Parallel circuit currents
Parallel connections between the phases of the two circuits, shown in figure5.2(b), will force the voltages to be the same in the two circuits. It will alsoallow for intrastator currents driven by the induced voltage difference to flowbetween the circuits. These currents do not contribute to the power deliveredby the generator but they do contribute to the resistive losses.
Another consequence of the parallel circuit currents is a reduction of thedifference in flux density between the two halves of the air gap and hence alowered UMP in equations (3.13) and (3.14).
5.2 Damper windings and rotor eccentricityDamper windings can reduce the UMP caused by rotor eccentricity [9] anddamper winding influence on UMP is discussed in article III. The measure-ments and simulations were performed with a relative eccentricity of 0.44, afield current of 15A and without the damper winding mounted.
ConfigurationsA production generator will typically have five to seven damper bars. Twoforms of damper winding connections are common. In the first configurationonly the damper bars within a pole are connected, through a short circuitingconnector or plate, and the only path for any interpole damper currents arethrough the contact between the damper bars and the iron components of therotor. This configuration is called a partial or a grill damper winding. The
other alternative is a complete damper winding where the damper bars of eachpole are directly connected to the bars of the adjacent poles at both ends of themachine through end ring segments. In the literature the expression amortis-seur winding is often used instead of damper winding.
17
7/29/2019 Ump Fulltext01
28/62
Four damper winding configurations were used in the measurements on theexperimental generator. Including the reference measurement of the UMP atstandstill the configurations were:
1) standstill.
2) without damper winding (WD).3) with a continuous damper winding with damper bar 2 only (D2-C).4) with a non-continuous damper winding with damper bars 1 and 3 (D13-
NC).5) with a continuous damper winding with damper bars 1 and 3 (D13-C).
Short theoryWhen a machine equipped with a partial damper winding is subject to static
rotor eccentricity the varying flux density in the air gap induces a voltage in thedamper bars. The resulting current creates an MMF which strives to counteractthe change in flux density; when the flux density is increasing the flux causedby the damper MMF is opposing the air gap flux, and conversely, when the airgap flux density is decreasing the damper winding induced flux is added to theair gap flux.
The magnitude of the induced damper winding voltage, UD is a function ofthe rate of change of the air gap flux density and therefore proportional to thederivative of the air gap permeance described by equation (3.8). With set to
zero in (3.8) the relationship is
UD =AMf0
0
sin()
(1 cos())2(5.4)
where A is the damper winding area and it has been assumed that the fluxdensity is uniform over the width of the damper winding. In section 7.5 themeasured damper winding voltage is fitted to this equation.
Damper winding currentsThe equations above are correct for an idealised situation with a partial damperwinding and does not account for stator slot ripples or the effect of eddy cur-rents on the air gap flux density. To obtain more information about the effectof damper windings on UMP and the magnitude of the currents in the damperwinding both simulations and measurements of different damper winding con-figurations were performed.
On the experimental generator it is possible to choose between one, two andthree damper bars and to select whether or not to connect the damper bars onadjacent poles. The damper bar positions in the pole shoe on the experimentalgenerator can be seen in figure 5.3.
18
7/29/2019 Ump Fulltext01
29/62
1.0s 1.2s
Bar 1 Bar 2 Bar 3
Rotation
Figure 5.3. The position of the damper bars seen from above.
5.3 Field winding interturn short circuit
In article IV the UMP was caused not by rotor eccentricity but by a shortcircuit between one or more field winding turns on one of the poles. Ageingand mechanical wear of the field winding insulation can result in electricalcontact between field winding turns. The result is a rotating unbalance whichcan excite vibrations in the generator.
ConfigurationsVarious degrees of field winding interturn short circuits were measured. The
flux in a magnetic circuit is
=NI
R(5.5)
where N is the number of turns in the coil, I the current in the coil and R isthe reluctance of the magnetic circuit.
In the FEM simulations the reduction in was accomplished through areduction of N in the field winding for one pole, while on the experimentalgenerator, some of current, I, was shunted past one pole through a resistorparallel with the pole field winding. The configurations are REF, 4.3, 9.5,
17 and 30 where REF is the unmodified rotor and the numbers refer to thereduction in ampere turns in percent on pole eleven.
Unbalanced magnetic pull from field winding short circuitThe effect on the air gap flux is described in chapter 7 and in article IV. A UMPwhich rotates with the mechanical frequency caused by the reduced flux fromthe affected pole is detected. Another observation is that the unaffected poles
with the same polarity as the damaged pole experience an increase in fluxdensity. The effect is seen in the measurements, the finite element calculationsand in the simple magnetic circuit representation of the generator depicted infigure 2.2.
19
7/29/2019 Ump Fulltext01
30/62
5.4 Standstill frequency response testIn article V a standstill frequency response, SSFR, test on the experimentalgenerator is performed to identify the machine transfer functions. Unlike thestandard SSFR test, which is described in [10], the study presented in article
V includes measurements of damper winding currents and voltages. This re-quired a redefinition of the damper equivalent circuit from the form represent-ing the physical damper winding to a dq representation. The new equivalentwinding is described in section II of article V.
ConfigurationsIn the test a voltage with varying frequency is applied to two of the statorphases while the third is left open. The result is a magnetic flux in the air gapwhich fluctuates in amplitude but does not rotate. The test is performed withthe rotor in the direct, d, and the quadrature, q, axis positions.
The test was performed with the applied voltage varying in frequency from5mHz to 300Hz. Complete and partial damper winding configurations wereused as well as a rotor without any damper winding.
It is stressed in [10] that the stator winding must keep a constant temperat-ure so the test is performed at current levels which are only a fraction of therated levels. Hence the obtained parameters do not include effects of magneticsaturation.
20
7/29/2019 Ump Fulltext01
31/62
6. Experimental generator
All articles included in this thesis contain measurements from the experi-mental generator Svante at the Division of electricity at Uppsala university.The construction of the generator represent a substantial part of the work be-hind this dissertation. Article I describes the construction of the generator.This chapter contains some additional information and includes work doneafter the publication of the article.
6.1 ConstructionThe construction process started with the donation of a twelve pole three phasevertical axis synchronous motor to the division.
Original motor
Other characteristics of the machine was that it had salient poles with solidpole shoes and that the rating was 185kW at 380V and a cos of 0.9. As anexperimental machine the motor was not very suitable in its original configur-ation since very few components were accessible for measurements and therewere no adjustment capabilities to modify parameters of interest. On the otherhand, the machine was in good condition and perfect as a base for a test setup.
A few modifications were made to the rotor and the stator of the motor toturn it into a generator more similar to hydropower production machines andto allow for the reconfigurations needed for various experiments. The rotor
and the stator were also individually suspended to allow for measurement ofthe magnetic pull in the air gap.
RotorTo reduce the risk of the rotor coming into contact with the stator during non-centered operation the average air gap was increased through a reduction of thepole body height. Given that there had to be room also during non-centered
operation for the air gap flux density sensors a longer air gap allowed for alarger relative eccentricity. It was also easier to implement certain relative ec-centricity on a larger air gap. Finally, a larger relative eccentricity increasedthe signal to noise ratio in some of the experiments that were planned.
21
7/29/2019 Ump Fulltext01
32/62
Laminated pole shoes
Laminated poles and rotor rings are standard on large hydropower generators[11]. On the experimental generator the original cast rotor was kept but, asthe pole bodies were shortened more than was required for the increased air
gap length, laminated pole shoes could be mounted. The laminated pole shoesserve to reduce the eddy currents in the pole face resulting from variations inthe air gap flux density. Under most circumstances the rotor poles are subjectto a smaller variation in the flux density than the stator, which sees a completereversal of the magnetic field every electric cycle, and therefore the laminationthickness is often thicker in the rotor than in the stator. The pole shoes of theexperimental generator were made of 2mm varnished sheet steel comparedto the 0.5 mm original steel of the stator. Figure 6.1 shows the shape of thelaminated pole shoe and figure 4 in article I shows the geometry of the pole
shoe mounted on the pole body.1.0s 1.2s
Rotor fluxsensor 1
Rotor fluxsensor 2
Rotation
Figure 6.1. Two air gap flux sensors was mounted on one of the poles. The signalswas transmitted to a stationary computer through the wireless module shown in figure6.9.
Damper winding
One of the main experiments the generator was built for was to investigate theeffect of a damper winding on the reduction UMP. The rotor was thereforeequipped with a removable and reconfigurable damper winding. A cross sec-tion of a pole body with the pole shoe and the damper bar slots is shown infigure 6.1. The damper bar slots were not distanced evenly but rather with theseparation of 1 and 1.2 stator slot pitches. This was done so that the effect ofdamper bar pitch on stator voltage total harmonic distortion, THD, could beinvestigated.
An example of a damper winding configuration is shown in figure 6.2. Heretwo damper bars, out of three possible, and the inter-pole connectors are used.The damper bars and the inter-pole connectors are all attached to a copper barat each end of the pole. The damper bars are insulated from the pole plates by
22
7/29/2019 Ump Fulltext01
33/62
Air gap
Field winding
Coil
Empty
damper slot
Damper barsInter-pole
connector
Rotation
Figure 6.2. Inter pole connectors as well as individual damper rods can be removed.
In the figure two out of three possible damper bars are mounted leaving one position
empty.
Pole
Rotation
ZCiC1iD1
+
U1
ZD
ZC ZA
ZA iD3
+
U3
ZD
ZC
ZC iC2
Figure 6.3. Circuit diagram of a section of a complete damper winding with two
damper bars per pole seen from the rotor shaft. The interpole connectors are repres-
ented by the impedances ZC. The direction of rotation is anticlockwise
heat shrinkable sleeves. Figure 6.3 contains the circuit diagram for a complete
damper winding with two damper bars.
Stator
The stator could be used as it was with the exception of a division of the
double layer concentric winding into two circuits to enable the parallel circuit
experiment described in article II .
Stator suspension
Adjustable rotor eccentricity and measurement of UMP required independent
suspension of the rotor and the stator. Either the rotor or the stator had to besuspended in a way that allowed measurement of the radial forces. A motor
and gear box had to be connected to the rotor so it would have been difficult
to suspend that part of the generator. It was decided that the vertical axis con-
23
7/29/2019 Ump Fulltext01
34/62
figuration of the machine would be kept and that stator should be suspendedin steel rods with ball joints at the ends.
Fx1
Fx2
Fy1 Fy2
Fr
Ft Fecc
R
y
x
Figure 6.4. The radial support of the stator seen from above.
One benefit of this solution was that keeping the axis vertical made it pos-
sible to reuse the old bearings, which were in good condition. Other itemsthat were used directly from the motor were the slip rings and the brushes. Anew attachment for the brushes spring mechanism was however built. Newhousings for the bearings also had to be made.
The geometry of the stator suspension is shown in figure 6.4. From themeasured forces it is also possible to calculate the mechanical torque on thegenerator.
Parallel circuits
The physical separation of the stator winding into two halves was performedby cutting the connectors between the coils instead of cutting the coils them-selves. With this solution the connectors required to alternate between one-and two-circuit operation was greatly reduced. The two possible stator wind-ing configurations are shown in figure 5.2.
Svante
Figure 6.5 shows a drawing of the resulting machine with a few of the com-ponents labelled. Worth noting is that the rotor is fixed and that the stator issuspended in such a way that it can be moved horizontally. All componentsseen in figure 6.5, except for the motor and gearbox, were designed from the
24
7/29/2019 Ump Fulltext01
35/62
geometry of the donated motor and the desired capabilities of the experimentalgenerator. Some of the data for the resulting machine are listed in table 6.1.
Longitudinalsupport barThrust/journalbearingBrushholder
Radialsupport bar
Journalbearing
Inductionmotor Gear
3.
5m
Figure 6.5. A structure was built to independently suspend the rotor and the stator.The figure shows how the rotor rests on a thrust bearing and stator is suspended inmetal bars. One horizontal beam in front of the stator has been removed from thedrawing for clarity.
Drive, motor and gear box
A 75kW induction motor and an inverter were used to power the generator.The motor had four poles whereas the generator had twelve and a 3:1 gearboxwas installed between the motor and the stator to give a suitable working speedfor the induction motor and its inverter during synchronous operation of thegenerator.
Magnetisation equipment
The field current can be supplied either a by purpose built rectifier or from acontrolled laboratory supply. The former was used in articles I and II .
25
7/29/2019 Ump Fulltext01
36/62
Table 6.1. Main parameters for the refitted test generator.
Frequency 50 HzNumber of pole pairs 6Speed 500 rpmSlots per pole and phase 3
Number of stator slots 108Coil pitch 9Stator inner diameter 725 mmStator length 303 mmAir gap length 8.4 mmPower of driving motor 75 kWRotor weight 900 kgStator weight 700 kg
Synchronisation equipment
Synchronised operation is the usual mode of operation for hydropower gener-ators and synchronisation equipment has been installed on the test setup to cre-ate realistic load conditions. Grid connected operation also enables researchinto generator-grid interaction.
A resistive load has also been installed to free the generator from constraintson for example voltage and rotational velocity during operation under load.The symbols in figure 6.6 are explained in table 6.2
EXC
INV
U, f
IM
SG
R
Synchrono-scope
400V
11kV
Tm,
156/400V
T1
T2
Figure 6.6. A schematic of the experimental generator, its drive and the connection tothe grid. The symbols are explained in table 6.2.
26
7/29/2019 Ump Fulltext01
37/62
Table 6.2. Symbols used in figure 6.6.
EXC Field winding exciterf Grid and generator frequency comparisonIM Induction motorINV Inverter
R Resistive loadSG Synchronous generatorT1 156 V to 400 V transformerT2 400 V to 11 kV transformerTm Mechanical torqueU Grid and generator voltage comparison mechanical frequency
6.2 MeasurementsThe experimental generator was built to study electromagnetic phenomenaas well as to validate the finite element software described in chapter 4. Toachieve this several types of sensors were mounted on both the stationary andthe rotating parts of the generator.
All measurements presented in articles IIV were collected as voltage sig-nals and stored at a rate of 10 000 samples per second. In total 28 signals wereused. Apart from an amplification of the strain gauge signals no filtering orother form of signal manipulation was performed on these signals.
Measurements were taken with differential channels and sensors were, whereverpossible, kept floating. Twisted pair signal cables were used.
In article V additional temporary measurement points were used. In the art-icle the amplification of the signals from the induced damper and field windingvoltages and currents and the data collection frequency were selected depend-ing on the frequency of the applied voltage.
Non-rotating measurementsNot all measurements below were available in every experiment, there was forexample no damper voltage signal available during damper current measure-ments, but all available measurements were recorded.
Stator air gap flux
Hall sensors were attached to two stator teeth 90 mechanical degrees apart.The sensors were powered with a constant current of 1mA. The angular sep-
aration of the sensors would cause them detect diff
erent levels of flux densityduring non-centered operation. Hall sensors were selected instead of searchcoils because of their ability to measure a constant flux level and because oftheir higher resolution [12].
27
7/29/2019 Ump Fulltext01
38/62
Stator voltages
The parallel circuit voltages could be measured separately during no load op-
eration with two stator circuits. Each phase was monitored and in total six
channels were dedicated for stator voltage measurements.
Stator currents
Three Hall effect current transducers were mounted to measure intrastator cur-
rents in each phase during parallel circuit operation. Another three transducers
were measuring the current from the generator during operation with load.
Stator forces
Strain gauges were used to measure the forces on the stator. They were posi-
tioned so that amplitude and direction of the UMP as well as the torque on the
generator could be calculated from the recorded data. Figure 6.4 contains aschematic of the geometry. Four strain gauges were mounted in a full Wheat-
stone bridge on each of the four rods. The signals for the four bridges were
then amplified individually before being stored together with the other meas-
urement data. The linearity of the strain gauges and the gain of the amplifiers
have been calibrated with known weights. Figure 6.7 contains a photograph
of one of the four rods used to fixate the stator radially. The positions of the
rods on the stator can be seen in figure 6.4.
Figure 6.7. The stator was secured radially with four 16 mm metal rods. Figure 6.4
shows the position of the rods. Each rod was equipped with four strain gauges in a
wheatstone bridge.
Rotating measurement equipment
Measurements collected on the rotor were transferred via wireless LAN to the
measuring computer.
Rotor air gap flux
Two Hall sensors, of the same type as used on the stator, were mounted on one
of the poles. At eccentric operation these sensors detect a varying flux densityover one mechanical revolution. They also detect the stator slot ripple in the
air gap flux density. The position of the rotor flux sensors are shown in figure
6.1.
28
7/29/2019 Ump Fulltext01
39/62
Damper winding voltage
With the damper winding replaced with open circuit conductors the voltage
induced in the damper bars was measured.
Damper winding currentsOn one pole current Hall sensors were placed around the damper bars and
interpole connectors. There were three sensors for the damper bars and two for
the interpole connectors. Figure 6.8 is a photograph of the housing in which
the currents sensors were mounted. The positions of the sensors coincide with
the current arrows in figure 6.3. ZC in the figure represent the removable
interpole connectors.
Figure 6.8. On one of the poles the damper bar and connector currents were measured
with Hall effect current transducers placed on the lower side of the rotor. A sensor is
mounted around each damper bar above the flat copper piece. The interpole connect-
ors, the two black cables, pass through two more current sensors.
Wireless data transfer
The data collected on the rotor was transferred through a WLAN connection tothe stationary measurement computer. To minimise the g-forces on the trans-
mitter and have a free line of sight from the aerial to the stationary computer it
was mounted directly on the top end of the rotor axle, as shown in figure 6.9.
29
7/29/2019 Ump Fulltext01
40/62
Figure 6.9. Data collected on the rotor was transferred wirelessly to the stationarylogging computer. The transmitter was placed on top of the rotor axle.
30
7/29/2019 Ump Fulltext01
41/62
7. Results
The results from the four research areas presented in chapter 5 can be foundin their respective articles IIV. This chapter contains the most interestingfindings from articles IIIV and some additional material not included in thearticles.
7.1 Magnetic saturationHydropower generators are normally operated at some degree of magneticsaturation but to avoid nonlinear effects from saturation on the measurementsall experiments were carried out at unsaturated levels of magnetic flux density.
One way of finding the magnetisation where the iron is starting to saturate isto plot the no-load voltage curve and see at which field current the relationshipbetween the terminal voltage and the field current ceases to be linear. Figure7.1 contains the line to line voltage of the experimental generator as a function
of the field current. To within measurement accuracy there was no detectablesaturation at field currents below 15 A and all experiments were performed ata field current of 15 A or lower. The correct voltage for synchronisation withthe grid, through the transformer shown in figure 6.6 was obtained at a fieldcurrent of 13A.
The no-load voltage curve was also used to get the correct shape for theBH curve used in the simulations. Due to the unknown materials of the ro-tor and the stator the terminal voltage and flux density of the simulations didnot initially agree with the measurements but from the measured BH curve
it was possible to obtain simulated values which corresponded well with themeasurements.
7.2 Parallel circuits and rotor eccentricitySection 5.1 described the parallel circuit configuration of the experimentalgenerator. It was also mentioned that in addition to FE simulations an analyt-ical permeance model was used to study the effect of parallel stator circuits.
This study was performed before the air gap flux density sensors had beenmounted on the experimental generator. The shortest air gap is positioned at180 degrees in the plots. For a more detailed presentation of the parallel circuitstudy please see article II.
31
7/29/2019 Ump Fulltext01
42/62
0 10 20 30 40 500
50
100
150
200
250
300
350
400
450
Field current [A]
Terminalvoltage[
V]
Figure 7.1. The no-load voltage. All experiments were carried out at a field current of15A or lower to avoid effects of magnetic saturation.
Effect on unbalanced magnetic pullWith the rotor displaced parallel to the symmetry line of separation of the twocircuits, the x axis in figure 5.1, the effects on UMP and terminal voltage werenegligible but when the stator was displaced perpendicular to the symmetryline, the y axis in figure 5.1, a large reduction of the UMP could be observed.The results are summarised in table 7.1. The effect of eddy currents are notincluded in either of the computer models so the UMP at standstill is the sameas for the configuration with one circuit.
Table 7.1. UMP reduction from parallel circuits with the eccentricity perpendicular
to the symmetry line.
Configuration Measured FEM Permeance model
Standstill Fy [N] 4250 - -
One circuit Fy [N] 3216 4115 4072
Two circuits Fy [N] 1345 1734 1774
Reduction relative to one circuit Fy [%] 58 58 56
Parallel circuit voltagesFigure 7.2 contains the voltages of the two parallel circuits when the rotoris displaced parallel to the y axis. Due to the difference in the average fluxdensity for the two circuits the induced circuit voltages are not equal. In the
32
7/29/2019 Ump Fulltext01
43/62
figure circuit one, with the on average shorter air gap, is represented by solidlines for phases a, b and c. The average rms voltage difference between the twocircuits is 19 percent. The simplified circuit voltages of equations (5.2) and(5.3) predicted a difference of 31 percent. Two explanations of the discrepancy
are the linearisation of the equations and an error in the measurement of therelative eccentricity used in the equations.
0 50 100 150 200 250 300 350100
50
0
50
100
Position (degrees)
UPH
[V]
Ua,1
Ub,1
Uc,1
Ua,2
Ub,2
Uc,2
Figure 7.2. The open circuit voltages measured at a relative eccentricity of 0.24 anda field current of 15A for the two stator circuits. Circuit one, represented by the solidlines, had an, on average, smaller air gap and hence larger total flux than circuit two.The result is a difference in the induced voltages in the two circuits. The layout of thecircuits are shown in figure 5.1.
Parallel circuit currentsThe induced voltage difference between the two halves of the stator windingwill, if the two winding sections are connected in parallel, drive intrastatorcurrents. The measured currents from such an experiment are shown in figure7.3. An effect of these currents is a reduction in the air gap flux density differ-
ence between the two halves of the stator. It is this reduction in flux densitydifference which causes the reduction of the UMP shown in table 7.1.During the experiment the generator was operated at no-load and, besides
reducing the UMP, these currents only result in resistive losses.
33
7/29/2019 Ump Fulltext01
44/62
0 50 100 150 200 250 300 350
60
40
20
0
20
40
60
Position (degrees)
I[A]
Ia
Ib
Ic
Figure 7.3. The parallel circuit currents driven by the difference in induced voltagebetween the two stator halves shown in figure 7.2. Measurements were taken at no-load.
7.3 Damper windings and rotor eccentricityArticle III describes the effect of different damper winding configurations on
UMP during rotor eccentricity. A description of the damper winding config-urations and the terminology used here can be found in section 5.2. Belowfollows a brief summary of the findings from the article. The shortest air gapis positioned at 180 degrees in the plots.
Effect on unbalanced magnetic pullThe resistance of the damper windings and the fraction of the pole face area
covered by the damper winding were, through simulations, found to have alarge influence on the reduction of the UMP from the damper winding. Withthe damper winding currents from the experimental generator the effect on theUMP was negligible. All experiments and simulations were carried out at fieldcurrent of 15A and a relative eccentricity of 0.44.
Table 7.2 contains the simulated forces resulting from the four configura-tions described in section 5.2. In these simulations of the experimental gen-erator the damper winding resistances were set to result in the same damperwinding currents as in the measurements.
As a comparison a 28 pole generator used in hydropower production, witha damper winding covering a large fraction of the pole face, was studied insimulations. The damper winding resistances were set to values calculatedfrom material resistivity only and no contact resistances were included, much
34
7/29/2019 Ump Fulltext01
45/62
as would be the case for a generator with a soldered damper winding. Theresulting large reduction in UMP can be seen in table 7.3. In the table it is alsopossible to see that a relatively large tangential force component, approxim-ately a quarter of the total remaining UMP, is introduced by the damper wind-
ing. In neither the experimental nor the production generator does the presenceof damper winding interpole connectors have a large influence on the UMP.The level of interpole connector currents in the experimental generator can befound in article III. In the production generator they were approximately anorder of magnitude less than the damper bar currents.
Table 7.2. Simulated UMP change on the experimental generator for the studied
damper winding configurations. The last column is the change in UMP compared to
the configuration WD. The notation WD, NC and C are described in section 5.2.
Configuration Fx [N] Fy [N] Fecc [N] Fecc [%]
WD 9174 -1 9174 -
D2-C 9192 35 9192 0.2
D13-NC 9156 139 9157 -0.2
D13C 9150 188 9152 -0.2
Table 7.3. Simulated UMP change on a 28 pole production generator for the studied
damper winding configurations. The last column is the change in UMP compared to
the configuration WD. D1-7 indicates that all seven damper bars per pole were used.
Configuration Fx [N] Fy [N] Fecc [N] Fecc [%]
WD 325800 300 325800 -
D1-7-NC 184600 49250 190670 -41.5
D1-7-C 184200 49000 190220 -41.6
7.4 Field winding interturn short circuitUnbalanced magnetic pull caused by field winding interturn short circuit, de-scribed in section 5.3 was studied in article IV. The experiments were per-formed with a centered rotor at a field current of 15 A. In the figures the indexREF indicates the magnetically symmetric rotor and the numbers the degreeof interturn short circuit on pole eleven.
Resulting unbalanced magnetic pullA winding interturn short circuit on one of the poles results in a reduced fluxdensity for the pole. Figure 7.4 contains the absolute air gap flux densities
35
7/29/2019 Ump Fulltext01
46/62
0 50 100 150 200 250 300 3500
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Position [deg]
Airgapfluxdensity
[T]
BREF
B9.5%
B30%
6 7 8 9 10 11 12 1 2 3 4 5
Figure 7.4. The absolute air gap flux density from the magnetically symmetric rotorand two of the interturn short circuit configurations.
for the unmodified rotor and for two of the reductions in ampere turns on poleeleven, as measured on a stator tooth. In this figure an interesting feature is theeffect the damaged pole has on the flux density of the other poles. The reduc-
tion of ampere turns on pole eleven reduces the total number of ampere turnson the rotor and hence the total flux in the air gap. Gausss law for magnetismstates that the magnetic field is divergence free and hence the conservation offlux has to hold. In this situation the total flux for the odd numbered poles,with the same polarity as the affected pole, and the even numbered poles, hav-ing the opposite polarity, must be equal. All even numbered poles in figure7.4 exhibit a lowered flux density as expected from the reduction in the totalnumber of ampere turns. At the same time the odd numbered poles, except forpole eleven, experience an increase in flux density to compensate for some of
the reduction in the flux density of pole eleven.One result of the asymmetric flux is a pull on the rotor in the opposite
direction of the affected pole. Table 7.4 contains, among other quantities, themeasured and simulated unbalance forces caused by different levels of fieldwinding interturn short circuit on one of the twelve poles on the experimentalgenerator. The corresponding measured force orbits can be seen in figure 7.5.
Flux from a magnetic circuit modelA magnetic circuit calculation, derived from the circuit in fig 2.2, of the effectof interturn field winding short circuit on pole flux was also performed. Tosimplify the model the rotor and stator reluctances, per pole, were combined
36
7/29/2019 Ump Fulltext01
47/62
Table 7.4. Summary of the changes in ampere turns, flux for pole eleven, UMP and
terminal voltage relative to the magnetically symmetric rotor. Measured/simulated
results.
Configuration At[%] POLE11 [%] FU MP [N] U[%]
4.3 4.28 4.4/3.6 118/101 0.51/0.33
9.5 9.50 8.6/7.3 234/223 0.74/0.73
17 17.4 15.0/14.2 378/401 1.4/1.4
30 29.6 25.5/23.9 612/659 2.4/2.4
1000 500 0 500 10001000
500
0
500
1000
FX
[N]
FY
[N]
w
FREF
F4.3%
F9.5%
F17%
F
30%
Figure 7.5. The measured force orbits from various degrees of field winding interturnshort circuit on pole eleven. The black small orbit in the middle comes from mechan-ical and magnetic unbalances in the rotor of the experimental generator. Its maximumamplitude is approximately 70N.
into a single reluctance. The resulting fluxes, shown in figure 7.6 exhibit the
same pattern as described above for the flux density measurements. Article IVdescribes the magnetic circuit in more detail.
7.5 Results not included in the articlesDamper winding circuit parametersDamper winding resistance and inductance for the configuration with two
damper bars and no interpole connector, D13-NC, were determined from themeasured damper winding voltage and current.The stator slot ripple in the measurements was quite pronounced as can be
seen in figures 7.7 and 7.8. To avoid the problem of slot ripple the data for both
37
7/29/2019 Ump Fulltext01
48/62
6 7 8 9 10 11 12 1 2 3 4 5
5
10
15
Pole
Airgapfluxfor17%[mWb]
Magnetically
symmetric rotor
Figure 7.6. The pole flux from the configuration with 17% interturn short circuit onpole eleven calculated the magnetic circuit. The red line represents the pole flux forthe rotor without any parallel resistor on pole eleven.
voltage and current was fitted to the function shown in equation 5.4 before the
circuit parameters were estimated. The fitted data was then used to find RDand LD in
UD =RDID +LDdIDdt
. (7.1)
where UD and ID are the fitted voltage and current. The damper winding resist-ance, RD, was found to be 2.3m. As a comparison, the resistance calculatedfrom material resistivity only was 0.4m. Contact resistances between thedamper bars and the interbar connectors are believed to be the main cause
of the difference between the experimental and theoretical values. Regardingthe inductance, LD, some caution should be used in the interpretation of theobtained value of 4H. With a measured phase shift between voltage and cur-rent of 5.5 degrees, shown in figure 7.9, there is some room for measurementerror and at small angles the relationship between measured angle and the in-ductance is approximately linear. A fit of the measured data, where the slotripple is prominent, to equation 7.1 results in the same RD and an inductanceof 3.2H. The latter value for LD is most likely closer to the true value due tothe larger effect of the inductance at the slot frequencies.
The impact of the inductance is quite small at the mechanical frequency,8.3Hz, but at the slot ripple frequency of 900Hz the effect is large as can beseen from the difference in the relative slot ripple amplitude in the voltage infigure 7.7 and the current in figure 7.8.
38
7/29/2019 Ump Fulltext01
49/62
0 50 100 150 200 250 300 350
0.1
0.05
0
0.05
0.1
0.15
Position [deg]
Dampervoltage[
V]
Measured UD13
Fitted UD13
Figure 7.7. The measured and fitted damper voltage for the configuration with twodamper bars and no interpole connectors. The stator slot ripple in the damper voltageis more prominent in the part of the revolution where the air gap is small.
0 50 100 150 200 250 300 35050
0
50
Position [deg]
Dampercurrent[A]
Measured ID13Fitted I
D13
Figure 7.8. The measured and fitted damper current for the configuration with twodamper bars and no interpole connectors.
Flux density and damper voltage measurement verificationAs a verification of the measurement equipment the expected induced dampervoltage calculated from the measured variation in air gap flux density wascompared to the actual damper voltage. To avoid the noise of the stator slot
39
7/29/2019 Ump Fulltext01
50/62
0 50 100 150 200 250 300 350
0.1
0.05
0
0.05
0.1
Dampervoltage[V]
Position [deg]
40
20
0
20
40
Dampercurrent[A]
Fitted UD13
Fitted ID13
Figure 7.9. The fitted damper voltage and current for the configuration with twodamper bars and no interpole connectors. Note the different scales on the two y axes.
ripple the flux density was fitted to equation (3.11) before the expected inducedvoltage was calculated. Figure 7.10 contains the graphs of the measured andfitted rotor flux density. One thing worth noticing in the figure is how pro-nounced the stator slot ripple is near where the air gap is at its shortest, at 180degrees, compared to on the opposite side of the generator. From the rotor flux
the damper voltage the calculated as according to equation (5.4). The result-ing calculated damper voltage together with its measured value can be seen infigure 7.11.
40
7/29/2019 Ump Fulltext01
51/62
0 50 100 150 200 250 300 3500
0.1
0.2
0.3
0.4
0.5
Position [deg]
Rotorfluxdensity[T]
Measured BR1
Fitted BR1
Figure 7.10. Sensor one shown in 6.1 was used to measure the air gap flux density infront of one of the poles. Here both the measured flux density, containing significantslot ripple in the small air gap, and its fitted value can be seen.
0 50 100 150 200 250 300 350
0.1
0.05
0
0.05
0.1
0.15
Position [deg]
Damperv
oltage[V]
Fitted UD13
From flux, UD13
Figure 7.11. The damper voltage calculated from the variation in flux density in theair gap.
41
7/29/2019 Ump Fulltext01
52/62
7/29/2019 Ump Fulltext01
53/62
8. Summary of articles
In this chapter short summaries of the contents of the articles included in thedissertation are presented. The authors contribution to each article is also spe-cified.
Article IDesign and construction of a synchronous generator test setup
This paper contains a description of the functionality of the test generatorsetup. Characteristics of the generator as well as the installed measurementequipment are discussed. An initial test of the effect of damper winding con-figuration on voltage harmonics is included.
The paper was presented by the author, who is the main author of the paper,
at The XIX International Conference on Electrical Machines in Rome, Italy,
2010. It was also selected for inclusion in the IEEE Xplore database.
Article IIReduction of unbalanced magnetic pull in synchronous machines due to
parallel circuits
The first experiment on the test generator concerning rotor eccentricity is
described in this article. Measurements of the UMP and currents in the paral-lel circuits have been performed under static eccentricity with the stator wind-ing connected as one circuit or two parallel circuits. Experiments were doneat noload conditions. Two numerical studies of the force reduction are alsopresented, one using a finite element code and one using a permeance model.The reduction of UMP was strongly dependent on the direction of unbalancerelative to the line of separation of the stator circuits. Eddy currents induced inthe rotor during operation were found to reduce the standstill UMP with morethan 20 percent.
The author performed the measurements and the finite element simulationspresented and he is the main author of the article.The article was published in IEEE Transactions on Magnetics, vol. 47(12),
pp 4827-4833, December 2011
43
7/29/2019 Ump Fulltext01
54/62
Article IIIDamper winding influence on unbalanced magnetic pull in hydropower
generators with rotor eccentricity
Continuous and noncontinuous damper windings resulted in different damper
winding currents but their effect on the UMP was found to be very similar.Measurements of the UMP, damper bar currents and air gap flux density wereperformed on the experimental generator configured with static eccentricityunder no-load conditions. Finite element simulations of the different configur-ations detailing the damper winding induced flux changes are also presented.
The author performed the measurements and simulations presented and heis the main author of the article.
The article was submitted to IEEE Transactions on Magnetics, November
2012
Article IVDynamic unbalanced pull from field winding turn short circuits in hydro-
power generators
The UMP resulting from field winding interturn short circuits is invest-igated both experimentally and through numerical modelling. Flux densitymeasurements are performed on the experimental generator as well as on a 28
pole production generator. The occurrence of parallel stator branches com-bined with a field winding short circuit gives rise to circulating stator currentswhich are analysed. A field winding short circuit in one pole gives rise to anincrease in the flux in the other poles with the same polarity. This effect isexplained.
The author performed the measurements and simulations presented and heis the main author of the article.
The article was submitted to Electric Power Components and Systems, Feb-
ruary 2013
Article VStandstill frequency response test on a synchronous machine extended
with damper bar measurements
Standstill frequency response (SSFR) test data including damper windingmeasurements for three physically different damper winding configurations is
presented. A method to incorporate measured damper bar voltages and cur-rents in the SSFR analysis scheme is proposed. The validity of the generatormodels identified through the test is substantiated by comparison of the simu-lated and measured machine response to a drive torque step disturbance.
44
7/29/2019 Ump Fulltext01
55/62
The author contributed to the measurements performed for the study andthe development of the equivalent damper circuit.
Manuscript
45
7/29/2019 Ump Fulltext01
56/62
7/29/2019 Ump Fulltext01
57/62
9. Svensk sammanfatting
1878 kallade Aristide Bergs vattenkraften fr det vita kolet, fr. la houilleblanche, fr att visa p den mngd kraft som fanns ltt tillgnglig i snn ochisen i de franska alperna.
Vattenkraftutvinning i industriell skala i Sverige tog sin brjan 1910 i ochmed drifttagningen av de fyra frsta generatorerna i Olidans kraftverk i Gtalv. Det frsta strre vattenkraftverket i norra Sverige, Porjus, invigdes 1915.
Fram till att krnkraft brjade tas i bruk fr elproduktion i Sverige under
sjuttio- och ttiotalen stod vattenkraften fr den vervldigande delen av denelektricitet som producerades. Vattenkraften var och r en avgrande frutstt-ning fr den svenska industrin och vlfrden. I dag str vattenkraften fr 45procent, eller 65 TWh, av den elektricitet som anvnds i Sverige.1
ven om tekniken r vl utvecklad s finns det ett behov av en kad frst-else fr elektromagnetiska fenomen i synkrongeneratorer fr att undvika on-digt slitage, minskad livslngd och haverier. Oskerhet i elfrsrjningen svlsom hgre elektricitetspriser, saker som pverkar bde industrin och hushl-len, kan bli fljden av driftproblem inom vattenkraften.
Vid sidan om de ekonomiska konsekvenserna s br vattnet i vra redanutbyggda lvar av miljskl utnyttjas p ett optimalt stt i vl fungerande vat-tenkraftverk.
Asymmetriska magnetkrafter uppstr nr rotorn i en generator inte r cen-trerad i statorn eller nr det magnetiska fldet i luftgapet av ngon annan anled-ning inte r symmetriskt. Dessa krafter kan frkorta livslngden p generatoreller orsaka haveri genom kat slitage p lager och andra infstningskompo-nenter fr rotorn och statorn. Asymmetriskt luftgapsflde kan ocks induceradmpstavsstrmmar och cirkulerande strmmar i statorn vilket leder till re-
sistiva frluster och strre vrmeutveckling, vilket i sin tur medfr kade eko-nomiska frluster och ett kat termiskt ldrande hos maskinen.
Denna avhandling undersker i tre studier olika konsekvenser av magnetiskobalans hos synkrongeneratorer.
Arbetet avhandlingen grundas p har utfrts vid Avdelningen fr elektrici-tetslra vid Uppsala universitet. Resultat har inhmtats frn numerisk analys,analytiska modeller och mtningar p den experimentutrustning som byggtsav frfattaren.
Forskningen som presenteras hr r en del i ett Elektraprogram som he-
ter Utveckling av elektromagnetiska lastmodeller fr simulering av interak-tionen mellan rotor och stator i vattenkraftgeneratorer. Elektra finansieras av
1Elektricitet utgr en tredjedel av den energi som anvnds i Sverige [1].
47
7/29/2019 Ump Fulltext01
58/62
ABB, Trafikverket, Svenska Kraftnt och, genom Svensk energi, ett flertalproducent- och transmissionsfretag. Elektras syfte r att lngsiktigt strkaden svenska industrins konkurrenskraft och d i frsta hand elleverantrer,kraftindustri och tillverkande industri.
Avhandlingen inleds med en kort beskrivning av hur en synkrongenera-tor fungerar och hur rotorexcentricitet ger upphov till obalanskrafter. Drefterkommer i kapitel 4 en beskrivning av den finita elementmetod som anvnts idatorsimuleringarna. Det fljande kapitlet beskriver de fyra forskningsomr-dena parallella kretsar och rotorexcentricitet, dmplinding och rotorexcentrici-tet, fltlindningskortslutning och karakt